[TC13] Pentagon Clock - Wind-up Pendulum Clock

[Didumos69]

Pentagon Clock

This is a wind-up pendulum clock powered by two pullback motors. This clock is not true to real time but rather focuses on features such as a wind-up mechanism, pin wheel escapement and three concentric hands. It has been built completely in-system and can be reproduced without any fine-tuning, except for making all gears run smoothly of course (LXF-file here). The clock has a very consistent autonomy of 18 minutes. It features:

• A wind-up mechanism - A differential is used to prevent the hands from running counterclockwise when winding-up the clock. One of it's outputs is locked by default and slips during wind-up.
• Pin wheel escapement - Instead of using teeth, the escape wheel has three round pins that are stopped and released by a scissors-like anchor, which has been integrated in the pendulum.
• Three concentric hands - Planetary gears and a differential house make for a fictitious LEGO time system with 2 hours in a clock revolution, 5 minutes in an hour and 30 seconds in a minute. 

The pentagon-shaped clock face refers to the 5 minute hour. All features have been originally developed for this contest.

 

 

Spoiler

I'm working on a wind-up pendulum clock driven by a pullback motor for the TC13 contest. It needs two advanced mechanisms: an escapement mechanism to make the pendulum confine the movement of the clock and a ratchet mechanism to be able to wind-up the pullback without moving the clock. As for the first - the escapement mechanism - I have an idea worked out which is inspired by my compact 90 degree stepper.

In the image below, the upper knob wheel is attached to an axle driven by the pullback motor (geared up probably). The pendulum is attached to the same axle. Imagine the upper knob wheel wants to rotate clockwise, then the center knob wheel will want to rotate counterclockwise. The center knob wheel will push itself against the right stopper and will want to move to the left, but only if the pendulum moves to the left too, as it is integrated in the pendulum. When the pendulum moves far enough to the left, the center knob wheel will pass the right stopper, but the opposite knob of the same knob wheel will bump into the upper side of the left stopper. Now when the pendulum moves to the right, the center knob wheel will be able to pass the left stopper as well and the whole process will start over again. I hope .

I will post progress in this thread.

Edited January 31, 2018 by Didumos69

[Aventador2004]

Wish you luck, I wasn't planning on using mine for something like this yet. Good implementation.

[lcvisser]

Darn, I had the same idea... 

There's a lot of LEGO escapements online, see e.g. this collection by Ben van de Waal. I never built one, but appears that tuning has to be done carefully... I'm curious to see how your idea will work out. If it works as you hope, it would seem to be a very robust solution 

[Rudivdk]

Now that is something else than your average pullback model... I have no idea how you would be building this (I am totally not into clockworks), but I will be following with interest. Good luck in competition!

[mocbuild101]

I like that you're reusing your stepper, it will be interesting to see how well it works...

[Didumos69]

First attempt to make an escapement failed. The stepper is not capable of keeping the whole thing in motion.

Next attempt will be something like this. The frame is part of the pendulum.

[Victor Imaginator]

Another topic to follow) Wish you luck with clock.

[Leonard Goldstein]

That is great! I hope you put a cuckoo into it (maybe you can use one of these as a replacement: 2546, 98388, 92084).

Following 

[mocbuild101]

2 hours ago, Didumos69 said:

Next attempt will be something like this.

Looks like it should work, but whether it will be efficient enough, I don't know...

I once had a go at building a clock, and this is the mechanism I used:

The motor input is the green axle, and the pendulum attaches to the red plate.

[Didumos69]

Thanks for the support guys! This is quite outside my comfort zone, because the essential mechanisms can not be invented in LDD.

7 hours ago, mocbuild101 said:

I once had a go at building a clock, and this is the mechanism I used:

The motor input is the green axle, and the pendulum attaches to the red plate.

That's a nice solution! I like the simplicity and compactness of it. That's what I'm looking for too. Al though I will probably not be using studded parts.

[lcvisser]

Maybe something like this can be done with only studless?

 

[Didumos69]

11 hours ago, Ludo Visser said:

Maybe something like this can be done with only studless?

That is pretty ingenious! Especially the use of the diff. And you're right I'm looking for a studless solution.

I made a little progress. Instead of the single pin solution in my previous post, I used 2 bevel gears with 3 pins with tow ball. At least I have something oscillating.

 

Edited December 24, 2017 by Didumos69

[Davidz90]

Wow, nice one! If You can make it work with pendulum, go for it. The "oscillating mass" solution has a serious drawback that the period strongly depends on driving force, which is not constant with pullback.

I have used the escapement from the "mini clock" above multiple times, it is very robust.

However, for a long running time, escape wheel with many teeth is beneficial. For example, this one:

 

should work pretty well with pullback and it gives a very constant impulse to the pendulum, for extra accuracy.

 

BTW Merry Christmas :)

Edited December 24, 2017 by Davidz90

[Didumos69]

22 hours ago, Davidz90 said:

The "oscillating mass" solution has a serious drawback that the period strongly depends on driving force, which is not constant with pullback.

I noticed that too. I used it in a pendulum setup and it works realy nice, but it's still sensitive to a varying driving force. I know that the frequency of a pendulum does not rely on the mass of the pendulum, bu I think adding mass will make it less sensitive to a varying driving force. So what I'm thinking of is to turn the mechanism around and integrate the pullback motor in the pendulum. That would also add some weight.

This is what the escapement currently looks like.

And this is how it behaves with pendulum:

 

[Boulderer]

It’s got a really satisfying sound to it. 

[Davidz90]

6 hours ago, Didumos69 said:

I used it in a pendulum setup and it works realy nice, but it's still sensitive to a varying driving force. I know that the frequency of a pendulum does not rely on the mass of the pendulum, bu I think adding mass will make it less sensitive to a varying driving force.

Definitely. Ideally, a very heavy pendulum (like 1 kg or more) on a low loss suspension (for example, knife edge) gives the best accuracy. Low swing angle is also beneficial and cuts on the power demand.

 

I have noticed that in Your new video, there is still rubberband on the pendulum axle - I believe it is unnecesary - gravity only should suffice.

[Didumos69]

48 minutes ago, Davidz90 said:

I  have noticed that in Your new video, there is still rubberband on the pendulum axle - I believe it is unnecesary - gravity only should suffice.

You are right. It was a leftover of the original setup.

Good to read your confirmation on the mass assumption . Btw I think my escapement's swing angle is quite low.

Edited December 25, 2017 by Didumos69

[Gonkius]

Good luck to have a ratchet mechanism during wind-up - the clock will probably stop immediately during wind-up due to the friction everywhere. I ended up with a differential mechanism instead to not loose torque during wind-up.

If you need inspiration look into my automatic self-winding, and self-adjusting, clock. If you like to find solutions of your own, don't look...

https://www.eurobricks.com/forum/index.php?/forums/topic/113864-moc-fully-automatic-clock/

 

[Didumos69]

15 hours ago, Gonkius said:

Good luck to have a ratchet mechanism during wind-up - the clock will probably stop immediately during wind-up due to the friction everywhere. I ended up with a differential mechanism instead to not loose torque during wind-up.

If you need inspiration look into my automatic self-winding, and self-adjusting, clock. If you like to find solutions of your own, don't look...

https://www.eurobricks.com/forum/index.php?/forums/topic/113864-moc-fully-automatic-clock/

 

That's a very amazing piece of work! I'm indeed looking for my own solutions, but I sure can use some inspiration, so I watched your topic anyway. Thanks!

[Didumos69]

I finally had some time to work on the wind-up part. It took some trial and error to get the gear ratios and the length / weight of the pendulum right. With my first attempt I needed a very long pendulum with a lot of weight. At that point the autonomy was about 3 minutes. After some more gearing up (currently the pullback motor is geared up with a 20:1 ratio) and improved bracing (I once more experienced how important this is to make gears run smoothly (rattle happily)), the autonomy now is 12 minutes, but that really seems to be the max. When I add the hands of the clock, I might have to reduce the gear ratio, which will also reduce the autonomy. I hope I can keep it around 10 minutes. The wind-up mechanism relies on a differential: the rear output is blocked when the clock runs normally, when you wind-up the clock, the rear output slips.

It appears that when the pullback motor is wound-up completely, it starts off with a lot of power for a relatively short interval. Then it has a very long interval with relatively constant power and finally it ends with a short interval of little power. Another thing I noticed, is that when I wind-up the clock a little - not fully - while the clock runs, the clock accelerates shortly and quickly enters it's steady state again. I can't really explain why.

The whole setup currently works equally well with a short / heavy pendulum and with a long / light pendulum. For now I opted for the short pendulum, because eventually I prefer having a standing clock. Here's a video of the current state. It's quite a long video, but I wanted to give an impression of how constant the speed is.

The next step will be to add hands. I decided to allow my self to come up with a fictitious LEGO time system (e.g. a LEGO day might have 3 6 LEGO hours and a LEGO hour might have 9 minutes), as the main challenge for me is to have a model that shows how a clock works rather than having a clock that tells the exact time. I'm aiming for 3 hands, each with a different speed.

Edited December 29, 2017 by Didumos69

[nerdsforprez]

Wonderful job here...... Very satisfying video to watch.  This is something I will be following with a lot of interest....

[mocbuild101]

2 hours ago, Didumos69 said:

Another thing I noticed, is that when I wind-up the clock a little - not fully - while the clock runs, the clock accelerates shortly and quickly enters it's steady state again. I can't really explain why.

It is because the pullback motor has a mechanism inside it that allows you to wind it up at a higher gear ratio than when it's released, however, the mechanism doesn't switch gear ratios immediately, so there is always a short burst of high torque before it changes.

You can see this in action by winding up a motor while counting the revolutions, and then counting them again as you gradually release it - you will find there will be many more revolutions on release.

[Victor Imaginator]

Enjoyed the video. Very relaxing to watch how this thing work)

Weird idea: maybe placing pendulum on top of mechanism can help to get rid of this non-LEGO weight?

[BusterHaus]

4 hours ago, Didumos69 said:

 

The next step will be to add hands. I decided to allow my self to come up with a fictitious LEGO time system (e.g. a LEGO day might have 3 6 LEGO hours and a LEGO hour might have 9 minutes), as the main challenge for me is to have a model that shows how a clock works rather than having a clock that tells the exact time. I'm aiming for 3 hands, each with a different speed.

Very good progress so far. How synchronized is the current ticking with real time? Is it something you're aiming for, or is the Lego time system enough? 

[Didumos69]

4 hours ago, nerdsforprez said:

Wonderful job here...... Very satisfying video to watch.  This is something I will be following with a lot of interest....

45 minutes ago, Victor Imaginator said:

Enjoyed the video. Very relaxing to watch how this thing work)

Thanks! I like watching it too. The sound is addicting.

1 hour ago, mocbuild101 said:

It is because the pullback motor has a mechanism inside it that allows you to wind it up at a higher gear ratio than when it's released, however, the mechanism doesn't switch gear ratios immediately, so there is always a short burst of high torque before it changes.

You can see this in action by winding up a motor while counting the revolutions, and then counting them again as you gradually release it - you will find there will be many more revolutions on release.

Aha, that makes sense. At least I don't need to worry there is something weird with my setup.

46 minutes ago, Victor Imaginator said:

Weird idea: maybe placing pendulum on top of mechanism can help to get rid of this non-LEGO weight?

To get rid of the non-LEGO weight I will eventually make this a standing clock. I'm thinking of four large inclined pillars with the pendulum in the center.

Just now, BusterHaus said:

Very good progress so far. How synchronized is the current ticking with real time? Is it something you're aiming for, or is the Lego time system enough? 

Thanks. I decided to not synchronize this clock with real time, even though two clicks almost correspond to a second. I like challenges like the escapement, the wind-up mechanism, the optimal gear ratio and multiple concentric hands and I don't like being tied to specific RPM ratios for the hands.